Biswajeet Parija

Implemented advanced machine learning algorithms such as Random Forest and Gradient Boosting, achieving a 92% accuracy in proactively predicting adverse events. Integrated patient characteristics and treatment variables from historical trial data using logistic regression to enhance predictive capabilities. Developed machine learning models including ensemble methods like XGBoost, resulting in an 88% accuracy in predicting preliminary efficacy outcomes. Utilized machine learning methodologies such as regression models for model patient-reported outcomes with a high accuracy of 90%. Informed treatment decisions by predicting the impact of treatment on patient's life quality using time series analysis for dynamic predictions.

Objective Assessment: Evaluated diverse measures of patient's previous medication adherence through frequency analysis of a large administrative claims database, achieving an accuracy of 8%. Applied machine learning techniques to assess the relevance of various measures in predicting future medication adherence with a predictive accuracy of 88%. Model Building and Testing: Constructed logistic regression models incorporating baseline variables and various previous adherence measures, achieving an AHC-ROC score of 0.92 and utilizing cross-validation for robustness. Data-Driven Insights: Derived actionable insights from predictive models emphasizing the impact of demographic variables, comorbidities, and self-care services utilization and medication Use. Demonstrated a substantial improvement in prediction accuracy with an overall accuracy increase of 10% through the incorporation of advanced machine learning techniques and statistical tests. Communication and Reporting: Effectively communicated complex statistical findings to both technical and non-technical stakeholders. Presented actionable insights clearly and concisely, facilitating informed decision-making and inferences.

Co-founded a club for helping students across disciplines to learn computer programming and also compete in various inter-college programming competitions.